An Algorithm for Identity Theft Mitigation: Keypoint Signature Verification

Abstract

The existence of identity theft in society has become a major concern due to the effects it causes to those that are affected by it, more especially in the financial sector. Thus this thesis establishes the existence of identity theft issues in the financial sector loan sections and proposes an algorithm that addresses the mitigation processes of identity theft by having the signatures on the loan forms verified using the implementation of the proposed algorithm, then the results are compared with the human experts verification that are done on a daily basis. From the qualitative data collected from the four SACCOs presented indicate the 93% of the respondents knew that forgery of one’s signature in the SACCO exists and from the 93%, 95% of them had been victims of identity theft and 50% of them knew it after deductions were been made from their accounts. The algorithm was implemented in a prototype that was used to test the signatures that were corrected from various individuals that belonged to various SACCOs. The prototype had successfully verified 80.1% of the test signatures and as expected the highest results from the four Human experts verification of forged signature was 8.3% indicating that they had indicated more signatures as originals. The prototype thus recorded an accuracy of 91.4% and a precision of 60.0%. Key words: Algorithm, Identity Theft, Mitigation, Handwritten Signature, Signature Verification